Author: Uri Blass
Date: 17:53:33 01/02/05
Go up one level in this thread
On January 02, 2005 at 17:57:16, Ulysses Omycron wrote: >On January 02, 2005 at 17:19:52, Peter Berger wrote: > >>>Using Multiple Lines.. >>>Analysis by Shredder 8: >>> >>>1. ± (1.02): 1...Qd6 2.Re1 Qb6 3.Qc2 Bd7 4.Rad1 exd5 5.exd5 Rfe8 6.Bf3 e4 7.Nxe4 >>>Nxe4 8.Bxe4 Qxb2 9.Qxb2 Bc3 >>>2. ± (1.11): 1...exd5 2.exd5 Ne8 3.f3 Nd6 4.Bf2 Bf5 5.Qa4 Qe8 6.Qb4 Qd7 7.Rfe1 >>>Rfc8 8.Rad1 Bf8 9.Bc5 b6 10.Qh4 >>>3. ± (1.26): 1...Ne8 2.Qb3 exd5 3.Nxd5 Nd6 4.f3 Be6 5.Rad1 f5 6.Qa3 Rf7 >>>4. +- (1.41): 1...Nxe4 2.Nxe4 exd5 3.Nc5 f5 4.f3 b6 5.Nb3 Bb7 6.Rc1 Rc8 7.Qd3 e4 >>>8.Qd2 Rxc1 >>>5. +- (1.44): 1...Qc7 2.Rc1 Bd7 3.Qb3 Rfc8 4.Rfd1 Rab8 5.Qb4 a5 6.d6 axb4 7.dxc7 >>>Rxc7 8.Bxe5 bxc3 9.Bxc7 >>> >> >>Can we please first discuss this output and what it means? This will mean I will>state the obvious - please be patient with me, and not jump on me for that. >>>You see a depth in the engine window - let's assume the engine has just finished>an iteration here.>It tried to find the best move for the black pieces. It found out that this is>1. ..Qd6 which will leave white up by only 1.02 of a pawn, so about a pawn up>for white.>>Now the engine assumes that the move Qd6 is not existant and starts all over to>search this depth, to produce the second line. > >No, it doesn't, it seems it actually scored all the options, It scored all the option in the way that peter described(first searching the best option and later searching the best option when the best is ignored). so even if you put >the output to show all the moves, you may get (For example) that axb4 scores >5.34, a suicide move, but the scores would be used to sort the next iteration, >so if bxc3 scores 9.80 (The worst available move) it will be searched last in >the next iteration. (At least that's what REBEL 12.00.02 for MS-DOS seems to do >with the worses moves, but stills searches the most interesting [After the best >move] first). programs often do not search last the worst move in one best mode because they do not know which move is the worst. > >>The engine will always search for only the best move which will be the one that >>provides the highest eval - its algorithms don't allow it to do anything else - >>it doesn't know any better.>>When the engine searches for the best move for the first time at a given depth (>the first line), it has no clue what eval the other moves would provide, only >>that it is equal or worse to what it got for Qd6. When it searches the second >>line it has no clue which results it will get for 3 and 4, only that it should >>be worse if it didn't mess up. > >If you are right, then what ChessBase does is actually trickign the engine, >making it think that it already made the notasgood move, so it gets an score >because the best moves after it are searched and so and so. This is what >CM10thEd seems to do in Mentor Lines. Chessbase does nothing except telling the engine how many moves to search. Part of the programmers support that option by using the algorithm that peter suggest. Note that they may do things faster by not repeating the same iteration 3 times for 3 best move. The idea is that after you have 3 moves at depth 7 and starts depth 8 you get an exact score for all the 3 moves that you evaluated as best at depth 7 and for every additional move you search if it is better than the third best move and you calculate exact score only in case that it is better. > >> >>Do we agree with each other so far ? >> > >I may have no idea of what I'm talking about, please correct me. > >>Because if we do , I really don't get what all this discussion is about, and so >>it would be your task to explain now what all the computerchess guys are missing>about the great options the chessprogram provides when "searching multiple>lines" when it is about solving testpositions (or even worse real positions when>you wouldn't know what is the best move in the first place). > >They'd work at least to show you the score the engine gives to the best move, so >you know something is wrong with the engine, if it thinks the winning move is >bad, then you make a DPA to know why it thinks so (More on this below) > >>Let's assume in the above that 1. ...Qc7 is a brilliant move in reality that >>actually wins the game. Well, the engine has no clue about that. It thinks it >>will lose for black ( white up 1.44 pawns ) and that it sucks. Only if it is >>forced to forget about the best three moves , it will come up with it, and it >>still doesn't like it at all. >> >>What am I missing ? >> > >Deep position analysis let's you rearch a move to a given Ply/Time/etc. and get >other notasgoodmoves (Or moves you imput) to be Deep position analysed too, and >the next too, and the next till you stop it. > >The result? > >Well, if Qc7 is really the best move, eventually all the other moves after being >DPAed will drop their scores, and Qc7 will raise its score till it's the best >move, so you don't care actually for what move the engine is going to do, but >for what is the best move after DPAing it very deep. You have illusion that Deep position analysis can solve it. It can often not solve it because deep position analysis may miss the next move after Qc7 2.Rc1. The best move after 1....Qc7 2.Rc1 can be only number 10 in the list of moves. Deep position analysis is practically not better for playing strength than normal mode of the engine(otherwise programmers could tell the engine to use deep position analysis in games automatically). > >That's how solving chess is possible; and no, solving chess is NOT knowing all >the scores for all the possible positions on the board, but: "it's build on the >fact that if you never do e4 (For example) you don't need to know the Petroff >Defence, Ruy Lopez, 2, 3 and 4 nights variations, you just need to know what >it's the best defence against it, if it turns out that the french e4 e6 d4 d5 is >the best, then you have succefully prunned way more than 1000000000 moves and >1000000 variations already, since you are doing d4, you don't need to have them >on the book." All engines do exactly this by the alphabeta algorithm in their normal mode without deep position analysis. If they want to play 1.d4 and 1.e4 e6 is worse than 1.d4 for white based on their analysis they will not analyze other the petroff defence when they analyze 1.e4 Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.